WO2001064578A1

WO2001064578A1 - Conversion of hydrogen chloride gas with recovery of chlorine

Info

Publication number: WO2001064578A1
Application number: PCT/US2001/006305
Authority: WO
Inventors: Robert A. Rapp; Philip Vais
Original assignee: The Ohio State University
Priority date: 2000-02-29
Filing date: 2001-02-28
Publication date: 2001-09-07
Also published as: AU2001250784A1; WO2001064578A8

Abstract

A method and apparatus for the reactive removal of hydrogen chloride gas from a process stream with the recovery of molecular chlorine are disclosed. In a preferred method, hydrogen chloride gas is contacted with a high-surface-area-to-volume particulate iron oxide in the presence of hydrogen gas and the resultant iron chloride is further converted to iron oxide and chlorine. In a second preferred method, a flow of HCI and hydrogen gas are passed to a first bed of reactive oxide. A sensor determines when substantially all of the reactive material of the first bed is spent. A valve, activated by the sensor, diverts the flow of gas to a second bed while the reactive material in the first bed is regenerated and the chlorine gas is recovered. A preferred apparatus for the recovery of chlorine from hydrogen chloride gas comprises a plurality of reaction chambers each containing a metal oxide.

Description

Conversion of Hydrogen Chloride Gas With

Recovery of Chlorine

Technical Field of the Invention

The invention includes a method and apparatus directed toward the reactive removal of hydrogen chloride gas from a process stream with the possibility for recovery of molecular chlorine. The invention is directed to the removal of HCI gas from a gas stream and the subsequent conversion of the intermediate product to generate molecular chlorine gas. Background of the Invention Chlorine is extensively used by the chemical and petrochemical industries in the production and processing of organic feed stocks and products. Hydrogen chloride (HCI) and other volatile chlorides are environmentally hazardous waste products often formed during processing. Examples of such processes are the chlorination of benzene or toluene, the production of chlorine-bearing insecticides, the production of refrigerants and aerosols, and the production of silicones and polyurethanes.

In many of these chemical industries, the high cost of chlorine and the difficulties in dealing with HCI disposal have forced companies to attempt to recover the chlorine from the HCI waste product. Current recovery techniques in the chemical industry follow two basic principles: the direct oxidation of HCI with oxygen or an electrolytic

process.

Within the direct oxidation methods, there have been a number of different chlorine recovery techniques, which have all been based on the reaction:

2 HCI(g) + Vz O₂ (g) = H₂O(g) + Cl₂ (g)

The first process developed for the recovery of chlorine from HCI gas was called the Gossage invention (1836) which involved water washing of the gas in a coke- packed tower. The aqueous acid was then reacted with manganese dioxide to produce chlorine gas. This process was rather wasteful because large quantities of manganese dioxide were required and only 50% of the potential chlorine was recovered.

Later improvements on the process by Weldon were able to reduce the consumption of manganese. The aqueous hydrochloric acid solution is oxidized by manganese dioxide to produce manganese chloride, which is then reconverted into manganese dioxide by treatment with air in the presence of lime.

This process allowed the manganese dioxide to be regenerated and subsequently reused. In this way, the consumption of the costly manganese reactant was eliminated. Also, high final concentrations of nearly 90% chlorine gas were achieved in this way. The efficiency of the process was still poor; however, as about half of the chlorine content introduced as HCI was wasted in the formation of calcium chloride in the regeneration step. The final yield of chlorine based on HCI input was around 30%.

Various two-stage processes proposed since then have been based on the formation of a metal chloride by the reaction of HCI with a metal oxide and the subsequent regeneration of the oxide by oxidation of the metal chloride with air to produce chlorine. Several oxides of nickel, copper, magnesium, cobalt and iron have all been investigated. One of the most serious operating difficulties encountered with the use of these oxides is the high temperature required for the oxidation of the metal chlorides. Nickel and iron chlorides are vaporized at these temperatures and are lost by the flow of gases before they can be oxidized. Likewise, the equilibrium constants for the reoxidation reactions are not advantageous except for iron chloride. Magnesium oxide and magnesium chloride tenaciously retain water so that part of the chloride is converted to an oxychloride and HCI.

A competing recycling technique at this time was the Deacon process, which involved the oxidation of gaseous hydrogen chloride by air in the presence of a manganese or copper salt on an inert porous carrier as catalyst. The process followed the reaction:

4 HCI (g) + O₂ (g) = 2 Cl₂ (g)+ 2 H₂O (I) This process gained technical superiority over Weldon's technique in 1883 when Hasenclever introduced an additional processing step which produced pure HCI by dehydrating the aqueous acid. However, the process still suffered in several respects mainly due to the high reaction temperatures required to obtain a satisfactory rate of reaction with the known catalyst systems. First, copper chloride tends to volatilize resulting in a rapid decline in activity of the catalyst. Second, hot spots in the reaction bed were problematic due to the exothermic nature of the oxidation reaction. Finally, the equipment suffered from severe corrosion at the favored reaction temperature. The Deacon process was vastly improved in the early 1900s by the development of more effective catalysts and the use of fluidized bed techniques. As a result of fluidization, a uniform temperature could be maintained without difficulties arising from the exothermic oxidation reaction. The inability of the catalysts to achieve high activities even at the high temperatures of 400-650°C was overcome in the development of the Shell Process. This process used a new copper-containing catalyst, which made conversion rates of 77-80% achievable. This catalyst contained chlorides of the rare earth metals or scandium, yttrium, zirconium, thorium, and uranium and one or more alkali metal chlorides, with the best results occurring in conjunction with a silica gel substrate. Although the catalyst was in a semi-molten state, the reaction surface area was drastically increased, and lower reaction temperatures could be efficiently maintained to prevent the copper chloride vaporization.

By the 1970s, the Deacon process was modified to overcome the conversion limitations in the Kel-Chlor process by operating at lower temperatures with sulfuric acid to absorb water as it was formed. With the water removed, the Deacon reaction could proceed to complete conversion.

Recent works on direct oxidation methods for recovering chlorine from HCI have revisited the carrier catalyst processes initially explored by Weldon. In their 1990 patent, Minet, et al. devised a process in which copper oxide impregnated on an alumina or silica carrier reacts with HCI according to the reaction:

CuO + 2 HCI (g) = CuCI₂ +H₂O (g)

This chloridation reaction is maintained at reaction temperatures of 180-260°C and

essentially captures all of the chlorine from the gas stream. The water vapor is separated and readily condensed. According to this continuous process, the chloridated catalytic carrier material is then transported to a second reaction chamber where CuO is obtained according to the reaction:

CuCl₂ + ^ΛA O₂ (g) = CuO + Cl₂ (g) This reoxidation reaction, carried out at a temperature of 300-360°C, liberates chlorine

gas which is sent to an absorber-stripper system utilizing carbon tetrachloride. In this way, the chlorine gas is removed from the gas stream, condensed, and collected as a liquid chlorine product. But the process requires a complicated transport of solid

particles and fluidized beds. Apparently, these authors encountered difficulties in achieving conversion efficiencies greater than 60-70% because of an equilibrium conversion constraint in a single-stage catalytic oxidation process. To account for this problem, a two-stage process was devised in which the gases were passed through a chloridation reactor, an oxidation reactor, and then back into the chloridation reactor. Specifically, the HCI in the gas stream was partially reacted with copper oxide catalyst up to the thermodynamic limit. The chloridated catalyst was then passed to the oxidation reactor where a second stream of oxygen reacted with the CuCI₂ in the catalyst to release Cl₂ gas. This entire mixture of (now reoxidized) copper oxide catalyst and HCI and Cl₂ gases was then passed back into the first chloridation reactor to complete the HCI removal from the gas stream.

Summary of the Invention The invention includes a method and apparatus directed toward the conversion of HCI gas with the possibility for recovery of chlorine from the intermediate reaction product. The method comprises the steps of contacting a metal oxide with a gaseous stream containing HCI and H₂ so as to produce a metal chloride and water vapor which may be referred to as "chlorine fixing", and then reoxidizing the metal chloride to release chlorine, and optionally and preferably collecting or reusing the chlorine. The metal oxide can preferably be selected from iron oxide or manganese oxide. In the HCI- containing gas, H₂ is present in excess of the stoichiometric amount so as to assist in reduction of the oxide and to drive the reaction to completion, the excess H₂ being present preferably in an amount of 1 part in 50 of the stoichiometric amount, more preferably in an amount of 1 part in 10 of the stoichiometric amount and most preferably in an amount of 1 part in 5 of the stoichiometric amount, or greater. The apparatus for recovering the chlorine from hydrogen chloride gas may comprise a plurality of chambers containing a particulate metal oxide, with each of the chambers connected via a valve to a respective fluid conduit adapted to conduct a flow of gaseous fluid containing HCI and H₂ to each of the respective fluid conduits. The apparatus additionally comprises a sensor in each of the chambers to sense when a reaction approaches completion and to signal a valve, such as a relay-activated valve, to release the gaseous fluid to the next chamber.

The invention preferably includes an improved direct oxidation method which utilizes iron oxide catalyst material to capture the chlorine from HCI gas and hydrogen gas. The benefits of this invention are apparent in the thermodynamic calculations for this system in comparison to the previously established oxidation techniques.

Table 1 contains thermodynamic calculations for the manganese oxide catalyst system using the HSC Chemistry program (version 3.0) for PC Windows. Equilibrium partial pressure calculations were performed assuming a closed system at equilibrium at 1 atmosphere starting pressure. These calculations for manganese oxide are presented to illustrate the problems which arise for a system which behaves in a less than optimum manner.

Table 1. Thermodynamic Calculations for the Chloridation of Manganese Oxide and the Oxidation of Manganese Chloride

calculated based on a closed system with stoichiometric gas input

As can be seen from the calculated equilibrium constants in reactions 1 and 4 in Table 1 , both Mn₃O₄ and Mn₂O₃ react spontaneously with HCI to form manganese dichloride. However, chlorine gas is also released as an objectionable product, and the

equilibrium constants at 400°C are not adequate to capture essentially all of the HCI. The calculated equilibrium constants for reactions 2 and 5 show that the presence of hydrogen gas, even in an amount quite inferior to the HCI content, increases the equilibrium constant significantly, and the release of chlorine gas is virtually eliminated. Thus, for reactions 2 and 5 of Table 1, stoichiometric or higher hydrogen content in the HCI-containing gas stream greatly increases the fraction of HCI gas which is reacted with the catalyst. However, although reactions 2 or 5 serve as an efficient trap for HCI from a hydrogen-containing gas stream, the equilibrium constants for the chlorine release/regeneration by oxidation reactions 3 or 6 indicate inefficient reactions, with significant oxygen mixed with the product chlorine. Thus the reoxidation of the MnCI₂ product does not promise an efficient release of chlorine gas. In summary for the manganese oxide catalyst, without the presence or addition of hydrogen gas, reactions 1 and 4 show that molecular chlorine is released with water vapor as a product of the reaction. A secondary separation process would be required to trap or reclaim this chlorine gas. The presence or addition of hydrogen gas in excess of the stoichiometric amounts indicated in reactions 2 and 5 prevents the evolution of chlorine gas in the HCI capture process. For this reason, H₂:HCI ratios in excess of 1:4 for Mn₂O₃ and 1 :6 for Mn₃O₄ need to be maintained to prevent the formation of chlorine gas. In this way, the reaction efficiency of the manganese oxide catalyst is vastly increased by the combined reaction of HCI and hydrogen gases. As can be seen from the calculations in Table 1 for reactions 2 and 5, the equilibrium partial pressure of the unconverted HCI_. gas is very small after reaction with the manganese oxides.

The thermodynamics for the similar reactions involving iron oxides, shown in Table 2, are even more favorable for the capture of HCI to form a solid metal chloride. Furthermore, in contrast to the manganese oxide catalyst system, the reoxidation of the FeCI₂ solid product with oxygen results in the efficient release of Cl₂ gas containing negligible oxygen. The comparable reactions in Table 2 are numbered the same as the similar reactions in Table 1.

Table 2. Thermodynamic Calculations for Chloridation of Iron Oxide and the Oxidation of Iron Chloride

Similar to the manganese oxide reactions, reactions 1 and 4 in Table 2 show the reactions of iron oxides with HCI in the absence of excess hydrogen. Once again, molecular chlorine is evolved as a reaction product. The presence or addition of hydrogen in reactions 2 and 5 essentially prevents the formation of chlorine gas and also allows an even more favorable and essentially complete reaction with HCI gas (as can be seen from the respective equilibrium constants). Therefore, H₂: HCI ratios in excess of 1 :4 for Fe₂O₃ and 1 :6 for Fe₃O₄ need to be maintained in order to prevent the formation of chlorine during the HCI capture process.

The iron oxide material system is slightly more complicated than manganese oxide because of the possible formation of a second chloride (FeCI₃). This chloride is highly volatile at the expected chloridation reaction temperatures. Such FeCI₃ vapor would be a major disadvantage for the iron oxide reactant because of the continuous evaporative loss of the material, as well as the contamination of the purge gas with the gaseous chloride species. Fortunately, this disadvantage is circumvented with the presence of excess hydrogen. As can be seen in reaction 7 in Table 2, in the presence of excess hydrogen, FeCI₃ is not thermodynamically stable and the FeCI₃ vapor pressure over FeCI₂ would be low.

The second aspect of the catalyst reaction system which must be optimized is the reoxidation of the chloridated catalyst (the chloride salt) to release molecular chlorine. Although manganese oxide is effective in converting HCI to form MnCI₂ and the regeneration of the MnCI₂ by oxidation to form Mn₂O₃ does take place manganese oxide does not appear to be the optimum catalyst material. The equilibrium oxygen partial pressures in reactions 3 and 6 in Table 1 reveal that a high percentage of oxygen should be present after these reactions. Consequently, the exit gas from the reaction column during reoxidation by pure O₂ is a mixture of O₂ and Cl₂ gases. This mixture would then once again require additional processing for the separation of these gases. With respect to chlorine release upon the reoxidation, iron oxides should be far superior to manganese oxides. The equilibrium partial pressures of oxygen in reactions 3 and 6 in Table 2 indicate that the amount of oxygen in the exiting CI₂ gas stream would be negligible. Brief Description of the Drawings

Figure 1 is a drawing of the multi-chambered apparatus used for recovery of chlorine from hydrogen chloride gas, in accordance with one embodiment of the invention. Detailed Description of the Preferred Embodiments

For the particular application of degassing/fluxing of liquid aluminum illustrated in Figure 1 , a mixture of argon plus chlorine gas is bubbled into an aluminum melt in order to reduce the dissolved contents of hydrogen, alkali metals and alkaline earth metals. The melt holding tank is covered by a refractory lid to collect the evolving gas.

The preferred embodiment of the present invention is the use of a number of enclosed reaction columns containing pellets of fine powder of the iron oxide catalyst material, perhaps incorporated onto an inert ceramic support. Initially, the reaction columns would be heated externally to a reaction temperature of about 200-450°C, and

purged with an inert gas to remove oxygen from the column. Process gas mixtures evolving from the melt containing HCI and H₂ compositionally in excess of a critical H₂/HCI ratios described by reactions 2 and 5 of Table 2, would then be introduced into the column. If the evolving process gas contained less than the stoichiometric hydrogen amount, hydrogen would be added from an auxiliary supply. Previous experiments have determined that the reaction front progresses with transport-limited (reactant arrival) kinetics. The product water vapor is removed from the gas stream by a desiccant. As the exothermic reaction front progresses through the reaction column, its approximate location can be monitored by thermocouples positioned along the length of the reaction column. When an in-line sensor determines that the column is nearly spent, automated valves can redirect the flow of the gas mixture into another (oxidized) reaction column. In fact, the gas treatment loop of Figure 1 also achieves a recycling of argon, as well as reuse of chlorine.

The spent (chlorided) columns would then be reoxidized with oxygen at a

reaction temperature of about 200-450°C to release chlorine gas which could then be used directly as a reactant, as shown in Figure 1 , or otherwise compressed and bottled.

When in-line sensors detect that the chloride reaction column is nearly completely converted back into the oxide state, the oxygen flow would be redirected into another spent column or simply stopped, and the regenerated/oxidized reaction column would be ready to begin the cycle again.

Claims

What is claimed is:

1. A method of recovering chlorine from hydrogen chloride gas, said method comprising the steps:

(a) contacting a high surface area to volume particulate iron oxide with a gaseous stream containing HCI and H₂ so as to produce iron chloride and water vapor; the latter readily extracted using a desiccant, and

(b) converting said iron chloride to iron oxide and chlorine.

2. A method of recovering chlorine from hydrogen chloride gas, said method comprising the steps:

(a) contacting a particulate metal oxide with a gaseous stream containing HCI and H₂, said H₂ being present in an amount in excess of a stoichiometric amount, so as to produce a metal chloride and water vapor;

(b) converting by oxidation the said metal chloride to form metal oxide and chlorine.

3. A method according to claim 2 wherein said metal is iron.

4. A method according to claim 2 wherein said H₂ being present or added in an amount of at least 1 part in 50 in excess of said stoichiometric amount.

5. A method according to claim 2 wherein said H₂ being present or added in an amount of at least 1 part in 10 in excess of said stoichiometric amount.

6. A method according to claim 2 wherein said H₂ being present or added in an amount of at least 1 part in 5 in excess of said stoichiometric amount.

7. An apparatus for the recovery of chlorine from hydrogen chloride gas, said apparatus comprising: (a) a plurality of reaction chambers containing a particulate metal oxide, each of said chambers connected to a respective fluid conduit adapted to conduct a flow of a gaseous fluid containing HCI and H₂ thereto;

(b) valve adapted to direct said gaseous fluid containing HCI and H₂ selectively to each of said respective fluid conduits.

8. An apparatus according to claim 7 additionally comprising a sensor disposed in each of said reaction chambers to determine that substantial reaction between said particulate metal oxide and said HCI and H₂ has occurred and to issue a signal in response thereto, and a relay adapted to receive said signal and to cause said valve to direct said gaseous fluid containing HCI and H₂to a second of said reaction chambers.

9. A method of recovering chlorine from hydrogen chloride gas comprising the steps:

(a) contacting a first bed of a reactive oxide with a gas stream containing HCI and H₂ to produce a metal chloride and water vapor; (b) selectively switching said gas stream containing HCI and H₂ to a second reaction column by an automated valve upon detection by an in-line sensor that the reaction has consumed substantially all of the reactant material;

(c) contacting the spent (chlorided) reaction bed with a gas stream containing oxygen to evolve chlorine gas and simultaneously regenerate the catalyst material to the oxide state; and (d) selectively switching the oxygen-containing gas stream to another spent reaction column by automated valves upon detection by an in-line sensor that the reaction has nearly consumed all of the metal chloride material.

10. A method according to claim 9 wherein said HCI gas is contained in a mixture with an inert carrier gas with a certain minimal hydrogen content.

11. A method according to claim 9 wherein said oxygen-containing gas is a mixture of oxygen and an inert carrier gas.

12. A method according to claim 9 wherein said gas stream containing HCI and H₂ contains hydrogen in excess of a stoichiometric amount.

13. A method according to claim 9 wherein CO (g) is used as the reducing gas. At the intended reaction temperature range, CO (g) and H₂(g) are similar thermodynamically in their tendency to reduce metal oxides.

14. A method according to claim 13 wherein said reaction temperature is within a range wherein said CO(g) and said H₂(g) have substantially the same tendency to reduce metal oxides.